The MIT Sloan Alumni Club of Boston, The MIT Center For Digital Business, SIM Boston and MIT ILP, are hosting the eight annual MIT Sloan CIO Symposium today. The event is sold out with over 900 people registered. Here are my notes from the session: Academic Keynote Panel:
MIT’s Perspective: What Every CIO Should Know About the Future Impact of Digital Business. Participants included: Moderator: Jason Pontin, Editor-in-chief & Publisher, MIT Technology Review, panel: Prof. Erik Brynjolfsson,
Schussel Professor of Management, MIT Sloan, Dr. David Clark,
Senior Research Scientist, MIT CSAIL, Prof. Gregory McRae,
Professor Emeritus, MIT, and Prof. Alex (Sandy) Pentland,
Professor, MIT Media Lab. The notes are done real time so apologies for any typos
Erik’s work looks at how digital technology transforms business. They are looking at data driven analysis. In past two days more data collected by US business than in the total past. He said that most business revolutions started with better data measurement through out history.
The New York Times wrote about his work recently (see When There’s No Such Thing as Too Much Information, by Steve Lohr,). Erik and his colleagues, Lorin Hitt, a professor at the Wharton School of the University of Pennsylvania, and Heekyung Kim, a graduate student at M.I.T., looked at 179 large companies. They found that those that adopted “data-driven decision making” achieved productivity rates 5 to 6 percent higher than could be explained by other factors, including how much the companies invested in technology.
David works in computer science and artificial intelligence. He is looking at forces that will affect future of the Web. He is looking at policy issues about home use. Also internet security is a focus as a political issue. Third thing is what is happening with industries that are having a violent crash with Web (telcos. music, video). He noted recent news that 30% of Web traffic is movie and TV steaming.
Greg is looking at how to design next generation computer enviornment for financial services. Alex is working in the Media Lab. He is building tools to collect data about what is happening on the Web and inside the enterprise. He sees real financial benefits when this data is used. However, there are security and privacy issues and regulators are getting involved.
Erik said there are over 1 billion searches on Google and you can access the aggregated data for free at Google Trends and other sources. They did it and could better predict housing sales than the professional real estate associations. Alex said you make a lot of predictions based on people’s Web behavior, including what apps you will download next, your political views, etc. You can get a better cross selling results in a ten-fold magnitude.
Greg said that data storage is increasing at a 80 percent rate compounded annually. There are new ways for data mining to get useful information from this data. This is important for financial services, especially the unstructured data.
Alex said that call center communication patterns can reveal data to improve productivity. They did this with one firm. You can also increase prediction results for diabetes risk as it is a behavior driven disease.
Erik companies with online data can do experiments more easily. Amazon does hundreds daily. However, you do not have to be online. Harrah’s in Vegas uses a lot of data on its clients. They want everyone in the company to think in a data driven way.
Jason asked the panel about security. David said there are specific obligations for privacy. However, the bad guys are more interested in breaking into servers than looking at individual transactions. Alex added that one strategy is to move the risk to the individual in the form of informed consent, perhaps with some benefit for accepting the risk. David is concerned about what happens with physical disks are stolen from data centers and it is unclear what was on them.
One audience member asked about security. Erik said there will likely be a huge security breach that will cause a wake up. Then companies will have to invest much more in it. Alex said that data auditing to track its sources may help but will be costly. Some firms and the defense agencies do this now. There are various encryption methods.
One audience asked about her noise in the data to make things simpler. How do you filter it? Greg said the cheap storage helps as then you can work on metadata issues and filtering when storing it. David said that just looking at data without caution can be misleading. Erik said that signal to noise ratio is getting better. The issue now is less about getting answers is asking the right questions. Greg said you can develop algorithms on what are the factor that most impact results. Erik said you need people who have the business skills and the analytical skills and this is a rare combination. Alex said you need to take the signal to noise experts and train them on the business.
He who successfully mines the largest database wins.
Posted by: twitter.com/TimCohn | May 18, 2011 at 11:38 AM
I love the term Erik coined to describe the opportunity embedded in data - "NanoData". Too many people talk about "Big Data," which emphasizes the noise and not the signal. Nanodata really capture where the opportunity is.
Posted by: Tod Loofbourrow | May 18, 2011 at 11:45 AM
Tod - Thanks for reminding me of this. Bill
Posted by: bill Ives | May 18, 2011 at 09:15 PM